Maternal CENP-C restores centromere symmetry in mammalian zygotes to ensure proper chromosome segregation

Summary:

Across metazoan species, the centromere-specific histone variant CENP-A is essential for accurate chromosome segregation, yet its regulation during the mammalian parental-to-zygote transition is poorly understood. To address this, we generated a CENP-A-mScarlet mouse model that revealed sex-specific dynamics: mature sperm retain 10% of the CENP-A levels present in MII oocytes. However, this difference is resolved in zygotes prior to the first mitosis, using maternally inherited cytoplasmic CENP-A. Notably, the increase in CENP-A at paternal centromeres is independent of sensing CENP-A asymmetry or the presence of maternal chromosomes. Instead, CENP-A equalization relies on the asymmetric recruitment of maternal CENP-C to paternal centromeres. Depletion of maternal CENP-A decreases total CENP-A in both pronuclei without disrupting equalization. In contrast, reducing maternal CENP-C or disruption of its dimerization function impairs CENP-A equalization and chromosome segregation. Therefore, maternal CENP-C acts as a key epigenetic regulator that resets centromeric symmetry at fertilization to preserve genome integrity.

Graphical Abstract

Mammalian zygotes inherit asymmetric levels of CENP-A between parental centromeres, which are remodeled during the first cell cycle to equalize CENP-A levels prior to zygotic mitosis using maternally inherited CENP-A. Equalization relies on maternal CENP-C and reduced CENP-C impairs proper chromosome segregation during the first mitosis.

Introduction:

Genome replication and segregation are fundamental processes that safeguard organismal development. During mitosis and meiosis, centromeres organize spindle microtubules and kinetochores to ensure faithful chromosome segregation. Centromeres are epigenetically defined by the presence of the histone H3 variant CENP-A^1,2. CENP-A-containing nucleosomes are extremely stable during the cell cycle^3–5 due to interactions with CENP-B, CENP-C, CENP-N, and HJURP^6–10– components of the constitutive centromere-associated network (CCAN)^11,12. After cell division, pre-existing CENP-A directs new CENP-A deposition during telophase and early G1^4,5,13. This one-to-one mechanism of CENP-A nucleosome incorporation ensures the precise propagation of centromeric chromatin, preserving both genomic location and quantitative nucleosome count across mitotic cell divisions.

CENP-A deposition and maintenance are highly regulated across the cell cycle, requiring at least four core factors: MIS18α, MIS18β, the MIS18-binding protein 1 (MIS18BP1, also known as KNL2), and the Holliday Junction Recognition Protein (HJURP), along with several accessory proteins/regulators^14–16. Although the deposition machinery is present throughout the cell cycle, CENP-A deposition requires integrated signals from polo like kinase 1 (PLK1) and cyclin-dependent kinases (CDK 1&2). At the end of mitosis and early G1, PLK1 phosphorylation of MIS18BP1 and HJURP promotes their localization to centromeres, while high CDK1&2 kinase activity at all other timepoints phosphorylates and sequesters MIS18BP1 and HJURP away from centromeres^16–23. Consistently, MIS18 complex formation, HJURP recruitment to centromeres, and CENP-A incorporation are all reduced in the presence of a PLK1 inhibitor¹⁸, whereas inhibition of CDK1/2 leads to ectopic deposition of CENP-A in G2 and S-phase^22,23. Importantly, dysregulation of any of these pathways renders the genome vulnerable to aneuploidy and genome instability, a hallmark of aging and cancer²⁴.

Although mechanisms governing CENP-A dynamics and inheritance in mitotic cells have been defined in vitro, our understanding of centromere regulation and maintenance in vivo remains limited. Studies in Drosophila and plants suggest that centromeric CENP-A levels are dynamically regulated during gametogenesis, with levels decreasing transiently during meiosis, before being restored in post-meiotic cells^25–27. In mice, CENP-A levels in female germ cells are maintained without active CENP-A deposition as loss of either Cenpa or Mis18α in early prophase-I-arrested oocytes has no or subtle effects on centromere maintenance, oocyte maturation, or female fertility^28,29. These findings indicate that mouse oocytes do not require new CENP-A deposition during quiescence, and new transcription is not required to resume meiotic divisions or support embryonic development. However, Cenpa^+/− mothers reportedly have decreased live birth rates and produce F1 males, but not females, with reduced CENP-A levels at germline centromeres³⁰. These weakened centromeres are propagated trans-generationally and are reset in F2 embryos between the 1-to-4 cell stage when F1 males are mated to wildtype (WT) females, suggesting that this correction process relies on maternal Cenpa genotype³⁰. Together, mammalian centromeres utilize genetic and epigenetic mechanisms to maintain and correct aberrations to centromeric chromatin from one generation to the next.

Across species, the physical inheritance and requirement of CENP-A at centromeres is variable³¹. In Caenorhabditis elegans, CENP-A is neither inherited nor required for re-establishment of holocentric chromosomes in the early embryo³². Similarly, CenH3 in egg cells of Arabidopsis thaliana is also not inherited³³. In contrast, in Drosophila melanogaster, inheriting the CENP-A ortholog, CID, is necessary for embryonic development²⁶. In humans, case studies have shown that neocentromere strength and position can be paternally inherited^34–36. Finally, in mice, somatic CENP-A is faithfully propagated across divisions, but the extent and mechanism by which parental Cenpa transcripts and centromeric nucleosomes are required to re-establish functional centromeres in the totipotent embryo is unknown.

Here, by quantifying CENP-A dynamics during male and female gametogenesis, we show that centromeric CENP-A levels are elevated in germ cell precursors relative to somatic cells. However, CENP-A loss is differentially regulated between the sexes, leading to sperm retaining roughly a tenth of the CENP-A levels present in mature oocytes. Following fertilization, CENP-A levels increase and equalize between parental chromosomes using maternally inherited, cytoplasmic CENP-A. Analysis of androgenetic embryos reveals that CENP-A accumulation in the paternal pronucleus is an autonomously regulated program, independent of the maternal pronucleus or any CENP-A asymmetry sensing mechanism, suggesting that sperm chromatin state dictates the maximum amount of CENP-A deposited on centromeres. This equalization process requires dimerized, maternal CENP-C. While equalization of parental CENP-A levels is necessary, it is not sufficient for successful embryogenesis. Zygotes with insufficient CENP-A can compensate through elevated CENP-C accumulation, highlighting a maternal safeguard mechanism that ensures faithful chromosome segregation and genome stability at the onset of development.

Results:

Generation and validation of tagged Cenpa^mScarlet/+ mouse model

To monitor CENP-A dynamics during mouse development and gametogenesis, we tagged the endogenous Cenpa gene with a C-terminal enhanced red fluorescent protein mScarlet-I and V5 epitope tag; the mice hereafter are referred to as Cenpa^mScarlet/+ (Figure 1A). Adult Cenpa^mScarlet/+ males displayed no overt abnormalities, including testes-to-body weight ratios and sperm counts comparable to WT littermates (Figure S1A&B). However, intercrosses of Cenpa^mScarlet/+ mice (het × het) yielded reduced litter sizes and no Cenpa^{mScarlet/mScarlet} embryos (0/18 embryos), while WT (Cenpa^+/+, 22.2%) and heterozygous (Cenpa^mScarlet/+; 77.8%) embryos were recovered at near-Mendelian ratios (Table S1&S2). The homozygous lethality is likely due to loss of CENP-A’s C-terminal interactions with CENP-C, which is known to stabilize CENP-A nucleosomes^7,37. Consistent with this, salt fractionation experiments show that although the CENP-A-mScarlet is incorporated into chromatin, it is less stable than untagged CENP-A (Figure 1SC). Importantly, although CENP-A-mScarlet has reduced stability in vitro at high salt concentrations, in vivo CENP-A-mScarlet localizes to centromeres: it enriches at DAPI-dense chromocenters in both testes and ovarian follicles, colocalizes with CREST or total CENP-A, and accumulates at chromosome ends in spermatocyte spreads–consistent with the telocentric organization of murine chromosomes³⁸ (Figure S1D). Importantly, CENP-A-mScarlet escapes chromatin remodeling during spermiogenesis and embryogenesis, allowing for parent-of-origin-specific tracking of CENP-A dynamics during mammalian development.

Figure 1: Loss of CENP-A in male germ cells precedes histone-to-protamine exchange.

A) Schematic of the Cenpa-GS-mScarlet-i-V5 transgene (https://biorender.com/ewmqnxf). B) Immunoblot of H3 and kinetochore components in flow sorted spermatogonia (‘gonia), pachytene/diplotene spermatocytes (‘cytes), round spermatids (‘tids), and mature sperm from Cenpa^mScarlet/+ males. Alpha tubulin serves as a loading control. Blots were stripped and re-blotted for the indicated proteins in n=2 mice. C) Schematic overview of germ cell markers used to define germ cell stages: GFRα1 (undifferentiated spermatogonia), SOHLH1 (differentiating A/Intermediate), STRA8 (B spermatogonia/preleptotene), and γH2AX (mid–late pachytene) (https://biorender.com/ella0sc). D) Representative immunofluorescence images of whole-mount Cenpa^mScarlet/+ tubules with various germ cell markers (n=200 cells per cell type; n=3 mice; scale bar: 20 μm). E) Quantification of CENP-A-mScarlet fluorescence across germ cell stages, testicular somatic cells, and intestinal cells (n=200 germ cells and intestinal crypt cells; n=100 for Sertoli and Myoid cells; n=2 mice). Each dot is the sum of CENP-A-mScarlet intensity per cell. F) Immunoblots of CENP-B protein levels in flow sorted germ cells and mature sperm from C57Bl/6J males. Representative image from n=2 replicates from n=2 mice.

Reduction of CENP-A in the male germline precedes the histone-to-protamine exchange

During spermiogenesis, the male epigenome undergoes a near complete replacement of histones with small basic proteins called protamines^39,40. Although retention of CENP-A nucleosomes has been reported in mammals, invertebrates, and even plants^{25–27,41,42}, the dynamics of CENP-A in both male and female gametogenesis remain poorly understood (see Das et al.⁴³, Štiavnická et al.⁴⁴). To answer this question, we used flow cytometry to isolate germ cell populations from Cenpa^mScarlet/+ males. As expected, histone H3 levels are fairly constant between spermatogonia and meiotic spermatocytes but decrease in post-meiotic spermatids and are largely depleted in mature spermatozoa (Figure 1B and S1E). In contrast, a significant fraction of CENP-A-mScarlet is lost at the transition from spermatogonia to spermatocytes. Subsequently, a smaller but notable loss occurs again during the transition from spermatocytes to round spermatids, with no significant change between spermatids and mature sperm (Figure 1B and S1E). Similar dynamics were observed when assessing total CENP-A levels in Cenpa^mScarlet/+ germ cells across stages (Figure 1B and S1E). Overall, these data suggest that tagged and untagged CENP-A proteins behave similarly, and that the loss of CENP-A from paternal chromosomes occurs before the histone-to-protamine exchange.

To pinpoint when CENP-A loss is initiated, we performed whole-mount staining of seminiferous tubules using markers to resolve germ cell subtypes. Within tubules, undifferentiated ID4+ and/or GFRA+ spermatogonia (A_single, A_paired, and A_aligned; collectively A_undiff) form syncytial clones of 2–16 cells (Figure 1C)^45,46. The A_al (4–16 cell clones) differentiate into A1 spermatogonia, then undergo transit amplifying divisions to make Type-B spermatogonia (SOHLH1+, then STRA8+)^47–49, which subsequently enter meiosis (γH2AX+)⁵⁰ (Figure 1C). When comparing across stages, we find that CENP-A-mScarlet levels were markedly higher in undifferentiated spermatogonia than differentiating germ cells, testis somatic cells (Sertoli and myoid cells) or even intestinal crypt cells (Figure 1D&E). Within the A_undiff pool, CENP-A-mScarlet levels dropped by ~30% between ID4^high (self-renewing) vs. ID4^low (committed) subpopulations (Figure S1F–I)^46,51,52. Taken together, these findings indicate that CENP-A loss coincides with stem cell differentiation and precedes chromosome remodeling during meiotic prophase and spermiogenesis.

To assess whether the observed decrease in CENP-A during murine germ cell differentiation is a unique or conserved phenomenon, we turned to Drosophila melanogaster. We collected adult fly testes and quantified the total fluorescence intensity of CID, the Drosophila CENP-A ortholog, in both somatic hub cells and germ cells at increasing distances from the hub, corresponding to different stages of differentiation. Consistent with our findings in mouse, CID levels were elevated in germline stem cells compared to hub cells and progressively decreased during spermatogonia transit-amplification divisions (Figure S2A–C). This suggests that CENP-A accumulation followed by a differentiation-coupled reduction is a conserved feature of male germ cell differentiation in both mice and flies.

Centromere constituents change throughout spermatogenesis

Next, we examined whether CENP-A interacting proteins have similar dynamics during spermatogenesis. Briefly, we found that CENP-C levels increase transiently in spermatocytes, despite significant loss of CENP-A, but decrease in haploid spermatids (Figure 1B and S1E). In contrast, CENP-B, which binds to B-box DNA sequences in minor satellites^6,53, decreased from spermatogonia to spermatocytes but persisted in spermatids (Figure 1F and S1E). In mature sperm, both CENP-B and CENP-C were undetectable, consistent with earlier observations in bull and Drosophila sperm^26,42. For HJURP, the CENP-A-specific chaperone protein, we observed high levels in spermatogonia and spermatocytes which decreased in round spermatids. Unlike CENP-B and CENP-C, low levels of HJURP persisted in mature sperm (Figure 1B and S1E). Furthermore, co-immunoprecipitations from mouse (Figure S1J) and human sperm (Figure S1K) indicates that sperm-retained HJURP associates with CENP-A/H2B nucleosomes. Taken together, these data indicate that many centromere-associated proteins (including CENP-A) are remodeled during germ cell differentiation, and that sperm-retained HJURP either associates with or localizes near CENP-A-containing nucleosomes, suggesting that paternally inherited HJURP may contribute to centromere function after fertilization.

CENP-A density at centromeres increases throughout oogenesis

To monitor centromere chromatin dynamics in the female germline, we quantified CENP-A and CENP-C in OCT4-eGFP^54,55 ovarian sections from primordial germ cells (PGCs; E11.5) through differentiating oogonia (E13.5, E18.5, P2) to adult GV oocytes (Figure S3A). As in the male germline, early female cells–PGCs and oogonia–displayed higher CENP-A levels than neighboring somatic cells. However, centromeric CENP-A levels continued to rise from birth to adult GV oocytes (normalized for ploidy) (Figure S3B). At the same developmental stages, CENP-C levels followed similar patterns as CENP-A (Figure S3C). After meiotic resumption, CENP-A-mScarlet levels peaked in GV and MI oocytes and declined slightly at the MI–MII transition (Figure S3D&E), whereas CENP-C levels increased throughout meiosis (Figure S3F). Thus, female germ cells progressively accumulate CENP-A and CENP-C during maturation, but a modest loss of CENP-A occurs during meiosis I which is offset by increased CENP-C.

Germline centromeric remodeling is regulated in a sex-specific manner

Given the differences in CENP-A dynamics in the male and female germlines, we next compared CENP-A levels in mature sperm and oocytes. After ploidy correction, CENP-A levels in sperm are significantly lower, containing only ~17.5% of the CENP-A present in GV oocytes and ~10.6% of that in MII oocytes by Western (Figure 2A&B). This difference was confirmed by immunofluorescence in MII oocytes and mature sperm (Figure 2C). To determine whether these differences are inherited, we collected zygotes from Cenpa^mScarlet/+ males and females immediately post-fertilization. Consistent with gamete measurements, the paternal pronucleus has 10–15% of the CENP-A-mScarlet signal intensity present in the maternal pronucleus (Figure 2D&E). This asymmetry is also evident in human gametes (~10% retention in sperm; Figure 2F) and, to a lesser degree, in Drosophila, where sperm retains ~60% of oocyte CID-Dendra2 (Figure 2G&H). The less pronounced difference in Drosophila gametes likely reflects the known post-meiotic deposition of CENP-A in males^25,26. Collectively, these findings highlight a conserved, yet species-tuned, difference in centromere composition between male and female gametes.

Figure 2: CENP-A asymmetry between oocytes and sperm is conserved across flies, mice, and humans.

A) Immunoblot of CENP-A and CENP-C in n=450 GV or MII oocytes and increasing sperm inputs. Membranes were stripped and re-probed with the indicated antibodies. Representative of two blots using n=12 females and n=2 males. B) Quantification of band intensities from (A), data normalized for ploidy, cellular input, and to GV protein levels. C) Quantification of total CENP-A immunofluorescence in MII oocytes and mature sperm from C57BL/6J mice (n=32 oocytes from n=6 females and n=600 sperm from n=2 males). D) Representative images of CENP-A-mScarlet fluorescence inherited from Cenpa^mScarlet/+ males and females. Data is from n=23 zygotes, n=2 technical replicates. Scale bar: 20 μm, 10 μm in inset. E) Quantification of CENP-A-mScarlet fluorescence in the maternal and paternal pronuclei of zygotes shown in panel (D). F) Quantification of total CENP-A immunofluorescence in human oocytes and sperm; n=5 oocytes and n=49 sperm. Data are from n=3 female donors and n=3 male donors and are normalized for ploidy. G) Representative images of Drosophila gametes with endogenous CID-Dendra2 fluorescence (red) and Hoechst (green); scale bars: oocyte = 10 μm, sperm = 1 μm. H) Quantification of CID-Dendra2 fluorescence from Drosophila gametes in (G) (n=62 oocytes, 94 sperm). ****p < 0.0001.

Paternal CENP-A-containing nucleosomes are inherited and persist in embryos

Using in vitro fertilization (IVF) with WT females and Cenpa^mScarlet/+ males, we next tracked paternal CENP-A-mScarlet dynamics during the first embryonic cell cycle, which can be grouped into five pronuclear stages (PN1 to PN5) based on the size and relative distance between the maternal and paternal pronuclei⁵⁶ (Figure 3A). Paternal CENP-A-Scarlet nucleosomes were detected in decondensing sperm (PN0), remained visible through PN1–PN5, and were retained in 2-cell blastomeres (Figure 3B). The persistence of tagged CENP-A in 2-cell embryos indicates that paternal CENP-A nucleosomes escape genome reorganization and DNA replication, contributing to embryonic centromeric chromatin.

Figure 3: Maternal- and paternal-derived CENP-A nucleosomes are inherited intergenerationally.

A) Schematic of zygotic pronuclear (PN) stages. B) CENP-A-mScarlet fluorescence in IVF zygotes from C57Bl/6J females × Cenpa^mScarlet/+ males at indicated PN stages (https://biorender.com/a1ojl2y). Maternal and paternal pronuclei were distinguished by size and position relative to polar body (n=4 males; 4 IVF/immunostaining replicates). Scale bars: 20 μm (main) and 10 μm (insets). C) Maternal CENP-A-mScarlet fluorescence across PN stages (n=12 females; EdU marks S-phase entry). Scale bars: 50 μm (main) and 20 μm (insets). D) Percent of zygotes with maternal CENP-A-mScarlet in one or both pronuclei at each PN stage. E) Percent of EdU positive or negative zygotes at each PN stage. F) Total CENP-A immunofluorescence in maternal vs. paternal pronuclei at PN5 (n = 25 zygotes); each dot represents summed puncta, normalized to maternal pronuclei. G) Comparison of maternally vs. paternally derived CENP-A-mScarlet levels at 8 hours post-fertilization. Only zygotes with maternal CENP-A-mScarlet present in both pronuclei were included. Each dot represents the total CENP-A-mScarlet in a single pronucleus. n=119 for maternally inherited CENP-A-mScarlet in maternal pronucleus, n=103 for maternal CENP-A-mScarlet in the paternal pronucleus embryos, and n=35 for paternal CENP-A-mScarlet in the paternal pronucleus. *: p < 0.05, ****: p < 0.0001.

Parental centromeres begin to equalize prior to DNA replication in zygotes

To track maternal CENP-A in zygotes, we next performed IVF using Cenpa^mScarlet/+ oocytes and WT sperm. As expected, CENP-A-mScarlet is detected on MII chromosomes post-fertilization, persisting through all pronuclear stages, and into the 2-cell embryo (Figure 3C). Interestingly, we find that maternal CENP-A begins to accumulate on paternal centromeres before S-phase (PN2/3; Figure 3C–E). By PN3, most zygotes (77%) have maternal CENP-A-mScarlet in the paternal pronucleus, while a smaller fraction (23%) showed localization only in the maternal pronucleus (Figure 3D). In late-stage zygotes (PN5), total CENP-A levels nearly equalize between the parental pronuclei (paternal/maternal ratio = 0.87; Figure 3F). Moreover, quantification of mScarlet fluorescence in PN3–5 zygotes revealed that the majority of CENP-A present in the paternal pronucleus is maternally derived (Figure 3G). These results demonstrate that CENP-A equalization is largely achieved prior to the first cell division.

CENP-A equalization is independent of zygotic transcription, translation, and DNA replication

Oocytes contain a significant reservoir of mRNA and protein that supports embryonic development prior to zygotic genome activation (ZGA) but are not believed to carry a large pool of free cytosolic CENP-A²⁹. To test whether deposition of maternal CENP-A onto paternal centromeres requires zygotic transcription, translation, or DNA synthesis, we performed IVF with Cenpa^mScarlet/+ oocytes and WT sperm and applied specific inhibitors 1–1.5 hours post-fertilization. Blocking minor ZGA with 5,6-dichloro-1-β-D-ribofuranosyl-benzimidazole (DRB)⁵⁷, a reversible RNA polymerase II inhibitor, efficiently suppressed nascent RNA synthesis as measured by 5-ethynyl-uridine (EU) labeling⁵⁸, but did not alter cell cycle progression (Figure S4G) nor did it prevent maternal CENP-A-mScarlet from accumulating in the paternal pronucleus (Figure S4A&D). Therefore, neither zygotic transcription of Cenpa nor transcription-coupled nucleosome turnover are required for maternal CENP-A-mScarlet deposition. Further, inhibiting zygotic translation using cycloheximide (CHX) abolished O-propargyl-puromycin (OPP)⁵⁹ incorporation in CHX-treated zygotes but not controls, confirming the effectiveness of CHX (Figure S4B). Despite this, maternal CENP-A-mScarlet deposition onto paternal centromeres was unaffected, though a statistically significant decrease is observed in the maternal pronucleus (Figure S4B&E). This effect persisted even when the analysis was focused on PN3-stage embryos to control for subtle cell cycle differences (Figure S4H&J). Together, these results suggest that newly made CENP-A protein in zygotes may be needed to achieve steady-state CENP-A levels at least in the maternal pronucleus. Finally, inhibition of DNA synthesis with aphidicolin (Aph), a reversible DNA polymerase inhibitor⁶⁰, is confirmed by loss of 5-Ethynyl-2′-deoxyuridine (EdU) incorporation (Figure S4C). Again, blocking zygotic DNA replication did not alter maternal CENP-A-mScarlet levels in the paternal pronucleus nor pronuclear staging (Figure S3C, F, I). Thus, paternal incorporation of maternal CENP-A-mScarlet does not require zygotic transcription, translation, or DNA replication, although some contribution from de novo translation is possible.

CENP-A equalization relies on maternally inherited CENP-A

Since zygotic transcription and translation contributed little to the equalization of parental centromeres, we next tested whether pre-existing, chromatin-bound maternal CENP-A nucleosomes and/or inherited cytoplasmic CENP-A protein contributes to equalization. To test this, we photobleached CENP-A-mScarlet nuclear fluorescence in GV oocytes, in vitro matured them to MII, fertilized them with WT sperm, and tracked CENP-A-mScarlet and total CENP-A fluorescence in the resulting zygotes (Figure S5A). Photobleaching depleted ~80–90% of the CENP-A-mScarlet signal in the GV nuclei while preserving oocyte integrity and competency (Figure S5B&C; see methods). Importantly, we did not observe significant photorecovery after holding GV oocytes for either one or 24 hours, nor following their maturation to MII (Figure S5B&C). These findings suggest that oocytes exhibit minimal to no exchange of chromatin-bound CENP-A during GV arrest or meiotic maturation, consistent with previous works using Cenpa and Mis18α conditional knockouts during oocyte maturation^28,29.

Given these observations, we next performed IVF using photobleached or unbleached in vitro matured oocytes. The resulting zygotes were cultured in the presence of CHX and collected eight hours after fertilization. Fertilized embryos derived from photobleached oocytes showed normal total CENP-A levels (Figure S5D&E), but markedly reduced CENP-A-mScarlet at maternal centromeres (32% of WT levels; Figure S5D&F). Surprisingly, despite near complete photobleaching (about 90%) in GV and MII oocytes, we detected a notable amount of new CENP-A-mScarlet in the maternal pronucleus, suggesting that new CENP-A is incorporated onto maternal centromeres. In contrast, the paternal pronucleus accumulates about 50% of CENP-A-mScarlet observed in WT embryos, consistent with CENP-A deposition in the paternal genome coming from two sources: recycled maternal CENP-A and a free inherited cytoplasmic pool of CENP-A from oocytes.

Since the photobleaching experiments were performed in the presence of CHX, the reappearance of new CENP-A-mScarlet in zygotes must derive from an inherited cytoplasmic pool. To confirm this, we quantified Cenpa transcripts across developmental stages and detected RNA in GV, MII, zygote, 2-cell, and 4-cell embryos. Cenpa RNA levels decreased from GV oocytes to zygotes but peaked after ZGA (Figure S5G). Additionally, re-analysis of low-input ribosome profiling data (Ribo-lite)⁶¹ confirmed that Cenpa mRNA is actively translated during oocyte maturation and early embryogenesis (Figure S5H). Taken together, our data suggests that increasing CENP-A levels at both parental centromeres likely relies on maternally provided CENP-A protein.

CENP-A accumulation in zygotes does not require sensing CENP-A asymmetry or maternal chromatin

Photobleaching experiments suggested that paternal centromeres acquire CENP-A from both a cytoplasmic pool and recycled maternal nuclear stores. To examine whether the recycled maternal nuclear pool is necessary for initiating CENP-A accumulation at paternal centromeres, we generated diploid zygotes using intracytoplasmic sperm injection (ICSI), as well as diploid androgenetic zygotes, which were created by injecting two sperm nuclei into enucleated oocytes (Figure 4A). In ICSI embryos, CENP-A levels at maternal and paternal centromeres equalize, similar to IVF-generated zygotes. Interestingly, in androgenetic zygotes the total CENP-A levels at paternal centromeres reach nearly identical levels to IVF or ICSI-derived embryos (Figure 4B). These findings demonstrate: (i) recycling of a maternal nuclear CENP-A pool is not essential for paternal loading or for directing CENP-A equalization, as the maternal cytoplasm alone is sufficient to support this process in the absence of a maternal pronucleus. (ii) Equalization does not require sensing parental asymmetry, as two sperm pronuclei with equally low CENP-A initiate the CENP-A deposition cascade and reach levels comparable to those of a normal paternal pronucleus, and (iii) since both sperm nuclei accumulate equivalent levels of CENP-A, this suggests that oocytes provide a non-limiting amount of CENP-A and CENP-A levels are restricted by properties of sperm-derived chromatin.

Figure 4: CENP-A incorporation in the paternal pronucleus is autonomously regulated.

A) Schematic of androgenetic embryo generation (https://biorender.com/fjua7jj). B) Quantification of total CENP-A immunofluorescence 16 hours after ICSI in control and androgenetic embryos (n=2 ICSI experiments). Each dot represents summed puncta per pronucleus; for fused androgenetic pronuclei, values were halved to normalize per genome. Fluorescence was normalized to the mean control maternal pronucleus (n=24 control maternal, 24 control paternal, and 38 androgenetic genomes). ns, not significant.

CDK1/2 and PLK1 regulate CENP-A deposition in zygotes

In somatic cells, CENP-A loading is constrained by CDK1/2 and PLK1 activity.^{16,18–20,22,23}. To test whether these kinases also regulate zygotic centromere remodeling, we treated zygotes with Flavopiridol (CDK1/2 inhibitor) or BI2536 (PLK1 inhibitor)^62,63. CDK inhibition (Flavo-treated) slows zygotic cell cycle progression in embryos and increases the proportion of PN2 zygotes with CENP-A-mScarlet incorporation in both pronuclei (Figure S6E&F). Notably, CDKi did not change CENP-A-mScarlet levels in the maternal pronucleus, but it did increase levels in the paternal pronucleus (Figure S6A&C). In contrast, PLK1 inhibition significantly reduced CENP-A-mScarlet levels in both pronuclei (Figure S6B&D), delayed cell cycle progression (Figure S6G), and reduced the proportion of zygotes with maternal CENP-A in both pronuclei (Figure S6H). Together, our data suggest that CDK1/2 and PLK1 activity, which normally suppress premature or ectopic CENP-A loading in somatic cells, also regulate CENP-A deposition and levels in zygotes.

CENP-C and MIS18BP1 are asymmetrically recruited to paternal centromeres in zygotes.

Next, to investigate how paternal centromeres achieve equalization, we examined the dynamics of centromere assembly factors in zygotes. In mitotic mammalian cells, CENP-C recruits the MIS18 complex and HJURP in early G1 to replenish CENP-A levels^7,8,64. However, in zygotes, paternal centromeres inherit only about 10% of the CENP-A levels present at maternal centromeres (Figure 2), requiring a 9–10-fold increase to achieve parity–an observation that challenges the classic one-to-one model of centromere inheritance described in mitotic cells. We therefore examined whether inner kinetochore components or CENP-A licensing/deposition machinery accumulate disproportionately on paternal centromeres. Two key CCAN proteins, CENP-B and CENP-C, are absent in sperm (Figure 1B and S1E) but are rapidly deposited onto the decondensing sperm genome in PN0-stage embryos, preceding CENP-A deposition (Figure 5A). Further, quantification of CENP-B-eGFP (via RNA microinjection; see methods) and CENP-C immunofluorescence reveals symmetric recruitment of CENP-B to both pronuclei, consistent with its DNA sequence-dependent recruitment mechanism^6,53, but an asymmetric recruitment of CENP-C to paternal centromeres (Figure 5B). This asymmetric enrichment of CENP-C at paternal centromeres continues to increase in later-stage zygotes (Figure 5B&C). Similarly, we also find that MIS18BP1 is preferentially enriched in the paternal pronucleus of zygotes, with this asymmetry again increasing in later-stage zygotes (Figure 5D&E). Together, these results suggest that preferential accumulation of both CENP-C and MIS18BP1 in the paternal pronucleus may contribute to CENP-A equalization at parental centromeres.

Figure 5: CENP-C and MIS18BP1 are asymmetrically recruited to paternal centromeres.

A) Representative immunofluorescence images of PN0–2 zygotes co-stained for CENP-C and maternal CENP-A-mScarlet (top) or total CENP-A and CENP-B-eGFP (bottom). Images are representative of n=2 IVF experiments using six Cenpa^mScarlet/+ or three CF-1 females and 1–2 B6D2F1/J male. Scale bars: 20 μm. B) Quantification of total CENP-C immunofluorescence or CENP-B-eGFP fluorescence in panel (A). Each dot is the sum of the total puncta in one pronucleus. Values are normalized to the maternal pronucleus (n=34 for CENP-C; n=8 for CENP-B-eGFP). C) Paternal/maternal ratios of CENP-C fluorescence in early (PN0–2, n=44) and late (PN3–5, n=90) zygotes. Each dot represents one zygote. The red dashed line represents a paternal/maternal ratio of 1. D) Representative images of MIS18BP1 localization across pronuclear stages (n=3 IVFs using n=7 females). Scale bars: 20 μm. E) Quantification of MIS18BP1 fluorescence in early (PN 1–3) and late (PN 4–5) zygotes from panel (D). Fluorescence was measured from a central z-slices and normalized to pronuclear area. Shown are paternal/maternal ratios; each dot is one zygote (n_early =10, n_late =10). Red dashed line = ratio of 1. p < 0.05; ns = not significant.

CENP-C directs CENP-A equalization and is critical for chromosome segregation during zygotic mitoses

Given the preferential accumulation of CENP-C in male pronuclei, we next tested whether CENP-C dosage influences CENP-A equalization. To this end, we knocked down maternally inherited CENP-C by microinjecting GV oocytes with a pool of Cenpc mRNA-targeting siRNAs. Considering CENP-C’s role in mediating interactions with other kinetochore components during chromosome segregation^1,65,66, we titrated the concentration of the siRNA pool to achieve a knockdown efficiency of about 50% (by RT-qPCR) to allow treated oocytes to progress through the remainder of meiosis and fertilization (Figure 6A). CENP-C knockdown preferentially impaired CENP-C’s recruitment to paternal centromeres (Figure 6B) and decreased the paternal-to-maternal ratio of CENP-A compared to control zygotes (Figure 6C–F). Although these embryos reached the 2-cell stage at normal frequencies (Figure 6H), they had a higher incidence of chromosome mis-segregation (Figure 6G&I), indicating that centromere equalization and chromosome fidelity are sensitive to CENP-C levels which is consistent with a CENP-C-dependent mechanism for CENP-A deposition^1,67,68.

Figure 6: CENP-A equalization and chromosome segregation rely on maternal, dimerized CENP-C.

A) RT-qPCR confirms efficient CENP-C knockdown in meiotic oocytes. Cenpc transcript levels were quantified from not injected, negative control siRNA (20–40 nM), and Cenpc siRNA (20–40 nM) groups (n= ~20 oocytes per replicate; n=2 biological replicates each with n=3 technical replicates. B) Paternal/maternal CENP-C fluorescence ratios in early zygotes (PN0–2) (n=3 replicates, using 3 CF-1 females and 1 B6D2F1/J each). The red dashed line denotes a ratio of 1. C) Representative PN4/5 zygotes from GV oocytes injected with Cenpc siRNA, control siRNA, or not injected controls stained for CENP-A (n=6 IVFs using n=3–4 CF-1 females and n=1 B6D2F1/J male each). Zygotes with paternal/maternal total CENP-A ratios above 5 were excluded from further analysis. Scale bar: 20 μm, 10 μm in inset. D) Quantification of paternal/maternal CENP-A ratios from (C). Each dot represents one zygote. Data: not injected (n=54), negative control (n=23), and siRNA-treated (n=35) zygotes from n=4 biological replicates. E-F) Scatterplots of maternal (x-axis) vs. paternal (y-axis) CENP-A fluorescence from (D) for not injected (E) and Cenpc siRNA-treated (F) zygotes. Linear regressions were performed with y-intercepts fixed at 0. Each dot = 1 zygote. G) Representative DAPI images of 2-cell embryos from conditions in (C), cultured for 24 hours post-fertilization. Scale bar: 20 μm. H) Proportion of embryos from (G) collected at each PN stage. “Stalled” indicates zygotes that failed to form a proper paternal pronucleus. Percentages were calculated by combining all embryos from all replicates. Data is representative of n=227 not injected, n=43 neg ctrl, and n=76 siRNA embryos, from at least n=3 replicates using n=5 CF-1 females and n=1 B6D2F1/J male each. I) Percentage of 2-cell embryos with at least one chromosome segregation defects from (G). Number of 2-cell embryos: not injected = 191, neg ctrl = 33, siRNA = 62. J) Rescue of CENP-A equalization by co-injection of siRNA-resistant Cenpc^WT RNA but not a dimerization-deficient mutant (Cenpc^Dimer). Paternal/maternal CENP-A ratios: not injected n=29, siRNA n=37, Cenpc^WT n=14, Cenpc^Dimer n=23. Data from n=3 replicates with n=5 CF-1 females and n=1 B6D2F1/J each. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant.

Because paternal centromeres start with low CENP-A yet accumulate disproportionately more CENP-C, we next examined whether CENP-C dimerization facilitates this enrichment. Indeed, prior studies have shown that CENP-C can dimerize or oligomerize, and mutations in residues important for dimerization impair CENP-C’s centromeric localization and its binding to HJURP^37,69,70. To test for a role of CENP-C dimerization during equalization, we knocked down maternal CENP-C in oocytes, but co-injected siRNA-resistant Cenpc mRNA encoding either the WT protein (Cenpc^WT) or a dimerization-deficient mutant mRNA (Cenpc^Dimer)^37,70. Notably, zygotes injected with the Cenpc^Dimer RNA failed to equalize CENP-A levels, phenocopying the siRNA-only condition, whereas Cenpc^WT RNA rescued the CENP-A equalization defect observed in Cenpc siRNA-injected zygotes (Figure 6J). This finding indicates that CENP-C dimerization is required for effective CENP-A equalization.

Finally, to assess whether CENP-A depletion in oocytes similarly impairs CENP-A equalization, we reduced the Cenpa RNA pool to approximately 20% of control levels (Figure 7A). This led to a global reduction in CENP-A at both parental centromeres (Figure 7B&C), yet equalization between pronuclei remained intact; suggesting that CENP-A equalization in zygotes relies on CENP-C levels. Curiously, in Cenpa siRNA-treated embryos, CENP-C levels were elevated in both pronuclei suggesting that additional CENP-C recruitment stabilizes weaker centromeres in zygotes (Figure 7D). Consistent with the knockdown experiments, zygotes from Cenpa^+/− females have reduced Cenpa transcript levels by RT-qPCR (Figure S7A), lower total centromeric CENP-A protein in both pronuclei, and increased CENP-C recruitment, particularly to paternal centromeres (Figure S7B&C). Although these effects were less pronounced in the Cenpa^+/− zygotes, this likely reflects the mixed genotype of oocytes generated from Cenpa^+/− females. Strikingly, embryos from either Cenpa siRNA-treated oocytes or Cenpa^+/– females showed no significant defects in 2-cell progression or chromosome segregation (Figure 7F&G), suggesting that reduced CENP-A levels can be tolerated if compensated for by increased CENP-C recruitment.

Figure 7: CENP-C compensates for acute CENP-A loss to preserve faithful chromosome segregation.

A) RT-qPCR quantification of Cenpa transcript levels in not injected, negative control siRNA (100–150 nM), and Cenpa siRNA–injected oocytes (100–150 nM) (n=6 biological replicates, n=10–20 oocytes each, 3 technical replicates). B) Representative PN3–PN5 zygotes (controls vs. Cenpa siRNA) stained for CENP-A (n=5 IVFs, 4–5 CF-1 females and n=1 B6D2F1/J male each). Scale bars: 20 μm. C) Quantification of total CENP-A immunofluorescence in maternal or paternal pronuclei from (B). Fluorescence measurements are normalized to the mean total maternal fluorescence in not injected oocytes. Shown above are the average paternal-to-maternal ratios for each sample. n=41 not injected zygotes, n=21 negative control zygotes, and n=26 siRNA zygotes. D) Quantification of total CENP-C in PN2/3 zygotes from Cenpa siRNA-treated zygotes generated as described for (C) from n=3 IVFs. n=54 not injected, n=23 negative control, and n=20 siRNA zygotes. E) Developmental outcomes at 24 hours post-fertilization for siRNA-treated embryos or embryos from Cenpa^+/– females. “Stalled” indicates zygotes that failed to form a paternal pronucleus. Data from all replicates consists of n=227 not injected, n=81 neg ctrl, and n=85 siRNA injected embryos (n=3 replicates with 5 CF-1 females and n=1 B6D2F1/J male each). Data also includes n=56 WT and n=56 Cenpa^+/− from n=2 replicates with 2–4 females for each genotype and n=2 B6D2F1/J males. Not injected embryos are the same as in Fig. 6H. F) Frequency of chromosome segregation defects in 2-cell embryos from (E): not injected n=191, negative control n=61, siRNA n=75, WT n=55, Cenpa^+/– n=56. G) Model for centromere remodeling and its impact on 2-cell embryo outcomes (https://biorender.com/lopqbuz). *: p < 0.05, **: p < 0.01, ***: p < 0.001, ****: p < 0.0001. ns indicates not significant.

Discussion:

Centromeres are epigenetically defined chromosomal regions marked by the histone H3 variant CENP-A and are essential for kinetochore assembly and faithful chromosome segregation. Despite this conserved function, CENP-A regulation is highly context-, cell type-, and species-dependent. Here we reveal dynamic, sex-specific control of centromere identity during mammalian gametogenesis and early embryogenesis and identify CENP-C as a key mediator of CENP-A equalization in zygotes.

In the male germline, CENP-A levels progressively decline during spermatogenesis, with the most pronounced drop occurring at the onset of differentiation and continuing through meiosis (Figure 1D&E and S1F–I). This reduction precedes global chromatin remodeling during spermiogenesis, revising earlier conclusions that CENP-A is stably retained throughout spermatogenesis^2,30,42,71. Our findings reveal a more nuanced view of CENP-A dynamics during early stages of male germ cell development. A similar pattern is observed in Drosophila, but unlike in mice and humans, CENP-A is redeposited post-meiosis^25,26. Given these differences, it is not surprising that a comparison of CENP-A levels between oocytes and sperm reveals that mammalian gametes have more dramatic CENP-A asymmetry than Drosophila (Figure 2). We speculate that the recovery in flies occurring in the male germline may be needed due to rapid cell divisions in Drosophila embryos, which does not provide sufficient time to overcome large centromeric asymmetry.

Sperm retain only about 10% of oocyte CENP-A, yet maintain centromere identity, suggesting a functional threshold required to epigenetically define a centromere. Supporting this, we observe limited variability in CENP-A intensity across individual sperm nuclei in humans and mice. In fact, the amount of CENP-A retained in sperm is comparable to the lowest levels tolerated in somatic cells before impaired kinetochore function is observed^72,73. Such low abundance likely necessitates specialized stabilization mechanisms. Notably, HJURP is the only CENP-A-associated protein tested retained in mature sperm (Figure 1B and S1E). Previous studies have shown that HJURP binding to CENP-A nucleosomes during passage of the replication fork maintains centromeric localization during mitotic divisions^10,74,75. We propose that paternally inherited HJURP may mediate a similar stabilizing effect on CENP-A during extensive chromatin remodeling of the paternal genome–both during spermatogenesis and in the zygote.

In contrast to males, female germ cells accumulate CENP-A from PGCs to GV oocytes (Figure S3B), with only a slight reduction from GV to MII (Figure S3E). Although centromeric CENP-A decreases slightly in MII oocytes, total CENP-A protein continues to rise (Figure 2A&B), consistent with a free cytoplasmic pool of CENP-A in oocytes. This conclusion is supported by our immunoblots of staged oocytes, photobleaching experiments, and reanalysis of published Ribo-lite data⁶¹ (Figure S5). This cytoplasmic pool of CENP-A is particularly intriguing because we show that CENP-A in mammalian oocytes is highly stable in GV and in vitro matured MII oocytes (Figure S5B&C). These findings align with prior genetic studies using conditional knockouts of Cenpa or Mis18α in growing oocytes showing that CENP-A levels are largely preserved during oocyte maturation, even in the absence of new deposition^28,29. Together, these results suggest that the cytoplasmic pool of CENP-A may serve as a stable reservoir of protein needed to support zygotic centromere remodeling.

These sex-specific dynamics during gametogenesis produce gametes with unequal CENP-A content (Figure 2). This asymmetry may be due to differences in CENP-A post-translational modifications (PTMs) in male and female germlines and/or, cell type-specific centromere-associated factors that can modify centromeric chromatin composition and structure. Indeed, CENP-A PTMs have been identified in mammalian cells, but their functions are not well resolved, and many of the PTMs identified in humans are not conserved even within the mammalian clade⁷⁶. Importantly, asymmetry in CENP-A is inherited by the mammalian zygote (Figure 2D&E). Although indirect methods have suggested that paternal nucleosomes are inherited and can affect early embryonic development^77–79, our study physically demonstrates and directly visualizes intergenerational epigenetic inheritance of CENP-A in mice. Conceivably, our observations of intergenerational CENP-A inheritance may extend to other histones retained in sperm, suggesting that paternal nucleosomes can serve as epigenetic memory carriers from father to offspring.

To resolve this asymmetry, zygotes rapidly recruit centromere assembly factors i.e. CENP-B and CENP-C post-fertilization, followed by MIS18BP1 (Figure 5). The enrichment of CENP-C at paternal centromeres is unexpected, given that its recruitment depends on CENP-A and CENP-B^1,6,7, which are present at equal or lower levels in the paternal genome compared to the maternal. This asymmetry in CENP-C suggests that centromere assembly is shaped by chromatin features intrinsic to the parental genome. Accordingly, androgenetic embryos accumulate CENP-A to levels indistinguishable from those of normal zygotes, indicating that CENP-A deposition, including the establishment of its upper limit, occurs independently of maternal centromeric input or asymmetry sensing.

Importantly, depletion of CENP-C in oocytes impairs CENP-A equalization and leads to chromosome segregation errors during the first mitosis (Figure 6D–I), underscoring an important role for maternal CENP-C in zygotic centromere remodeling. Unlike CENP-C knockdown, reducing maternally inherited CENP-A does not impair CENP-A equalization but does lower the overall CENP-A levels at both sets of parental centromeres. However, embryos with reduced CENP-A have a significantly higher amount of centromeric CENP-C levels (Figure 7D and S7C), suggesting a compensatory mechanism that may help stabilize centromeres. Others have previously proposed a similar mechanism in which centromeric CENP-C can be stabilized in a CENP-B-dependent manner, allowing cells to compensate for diminished levels of CENP-A⁶. A similar inverse relationship between CENP-A and CENP-C was also observed in spermatocytes, where decreasing CENP-A levels coincided with increased CENP-C accumulation, reinforcing the idea that upregulation of CENP-C supports weaker centromere function and allows chromosome segregation in both spermatocytes and zygotes (Figure 1B and S1E) These observations are consistent with prior work showing that chromosome segregation defects in CENP-A-depleted cell lines emerge only after concomitant loss of CENP-C^72,73, and with reports that Cenpc-null embryos exhibit earlier developmental arrest than Cenpa-null embryos^80,81.

Although the factors shaping sex-specific centromeric states remain to be defined, our findings point to asymmetric CENP-C recruitment as the driver of paternal CENP-A deposition in zygotes (Figure 7H). Proper regulation of centromeric chromatin during gametogenesis and early embryogenesis is critical, as disruptions can lead to aneuploidies, which are common in mammalian embryos^82,83,84. Notably, we also found that CENP-A and its assembly factors vary widely and are predominantly maternally inherited. Therefore, we propose that oocytes must maintain a sufficient pool of these proteins for successful embryogenesis, with variations in their levels potentially serving as biomarkers for oocyte quality and female fertility.

Limitations of the study:

All experiments using CENP-A-mScarlet were conducted with heterozygous Cenpa^mScarlet/+ mice, as Cenpa^{mScarlet/mScarlet} is embryonic lethal (see Table S1&S2), likely due to weakened interactions between tagged CENP-A and CCAN components³⁷. However, we do not expect use of this model to affect our conclusions for multiple reasons: 1) heterozygous animals are viable and fertile, and CENP-A-mScarlet faithfully localizes to centromeres and colocalizes with CREST. 2) Prior work shows that only about 50% of CENP-A nucleosomes need to engage kinetochore proteins for full centromere function⁸⁵. 3) Any effect of the tag would impact maternal and paternal genomes equally, therefore preserving the comparative dynamics we report. To ensure rigor, we validated key findings using both tagged and endogenous CENP-A and observed consistent localization and behavior in gametes and embryos, and all knockdowns were performed with WT gametes. Finally, we note that the reported chromosome segregation defects were solely assessed using DAPI which can both under- and over-estimate mis-segregation events. Whether these defects will persist into later cleavage divisions or impair implantation is not directly assessed.

Resource Availability:

Lead Contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Saher Sue Hammoud (hammou@umich.edu).

Materials Availability

All unique reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and Code Availability

All data reported in this paper will be shared by the lead contact upon request.

This paper does not report original code.

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR METHODS:

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mice

All experiments utilizing animals in this study were approved by the Institutional Animal Care and Use Committees of the University of Michigan (Protocols: PRO00006047, PRO00008135, PRO00010000, PRO00011691), John Hopkins University (Protocol: MO22A71), and Rutgers University (Protocol: 201702497) and were performed in accordance with the National Institutes of Health Guide for the care and use of laboratory animals. Mice were housed in an environment controlled for light (12 hours on/off) and temperature (21 to 23°C) with ad libitum access to water and food (Lab Diet #5008 for breeding mice and #5LOD for non-breeding animals).

Cenpa^mScarlet/+ knock-in mice were generated on the C57BL/6N background using CRISPR/Cas9-mediated genome editing by the University of Michigan’s Transgenic Animal Model Core⁸⁶. The sgRNA and donor oligo were designed as previously described⁸⁷, targeting the sequence 5’- AGAATCCGAGGCTTCGAGGGCGG-3’ in exon 4 of the Cenpa locus. The guide RNA target sequence was selected according to the on- and off-target scores provided by the web tool CRISPOR⁸⁷ (http://crispor.tefor.net) and proximity to the target site. Ribonucleoprotein (RNP) complexes were formed by mixing the sgRNA (2.5 ng/uL) with Cas9 protein (IDT, 5 ng/uL) in Opti-MEM (ThermoFisher) and incubating at 37°C for 10 minutes, at which time the donor oligo (IDT, 10 ng/uL) containing the intended transgene was added. The paternal pronucleus of zygotes generated from super-ovulated C57BL/6N females were microinjected with the RNP/donor oligo mix using a Femtojet 4i microinjector (Eppendorf). After injection, zygotes were moved to KSOM AA medium, matured to the 2-cell stage, and transferred to oviductal ampullas of pseudopregnant CD-1 females. Offspring were genotyped for the transgene insertion by extraction of genomic DNA from a small ear biopsy. Mouse lines from two founding CRISPR/Cas9 transgene insertions were maintained separately for Cenpa^mScarlet/+ on the C57BL/6J genetic background and utilized interchangeably in all applicable experiments. Transgenic and WT males were used for all experiments at least 8 weeks of age. Transgenic and WT females were used at 3–4 or 6–16 weeks of age.

The Cenpa^+/− mouse line was generated for a previous publication²⁹. Mice from this line were transferred to the University of Michigan animal facility where they are maintained on the C57Bl/6J genetic background. All females used for experiments were 6–16 weeks old.

Human sperm and oocyte samples

Deidentified human sperm and oocyte samples were collected from the University of Michigan Center for Reproductive Medicine. For samples collected for immunoprecipitation/immunoblots, the University of Michigan Institutional Review Board determined this study did not fit the definition of research involving human subjects (U.S. Department of Health & Human Services regulations at 45CFR46.102) because the research was intended to contribute to generalizable knowledge, the researchers did not interact with human subjects, nor obtained identifiable private information or identifiable biospecimens. Samples for immunofluorescence were collected under the University of Michigan Institutional Review Board (IRB00001996 approved 12/13/24, study ID HUM00264272). Oocytes were retrieved, fixed, and stored for staining on the same day according to the fixation protocol detailed below for mouse oocytes. As per the fertility clinics collection protocol, sperm from normospermic patients were processed and cryopreserved. Sperm samples were allowed to liquefy at room temperature (RT) in a sterile container for 30 minutes before being processed for either immunoprecipitation, immunofluorescence or snap frozen in liquid nitrogen and stored at −80°C for immunoblots.

Drosophila gamete samples

Drosophila melanogaster gametes were collected from CID-Dendra2 knock-in strains. This line was generated in collaboration with Fungene Inc. (Beijing, China) using CRISPR/Cas9 genome editing. Briefly, a Dendra2 fluorescent tag was knocked into CG6133 (cid) in-frame between codons 118 and 119, CID N term-Dendra2-CID C term⁸⁸. Mature oocytes and mature sperm were isolated from dissected adult Drosophila melanogaster females and males. Samples were fixed and imaged on a Lecia SPE confocal microscope for endogenous CID-Dendra2 fluorescence along with Hoechst staining.

METHOD DETAILS

Phenotypic assessment of Cenpa^mScarlet/+ mice

All phenotyping was performed in males 8–12 weeks of age. Body weight was recorded immediately after euthanasia, and testes were weighed within 5 minutes of dissection. Epididymal sperm counts and motility were obtained using a Makler chamber (Cooper Surgical SEF-MAKL). For each animal, three independent technical counts were performed from the same sample and averaged to obtain a single value; three biological replicates (three mice) were analyzed per founder line.

Fertility assessment of Cenpa^mScarlet/+ intercrosses:

Timed matings were set up for Cenpa^mScarlet/+ x Cenpa^mScarlet/+ mice to collect embryonic samples to assess homozygous lethality. Females were checked for copulatory plugs every morning after pairing with a male. After a plug was noted, females were separated into a different cage until the time of collection. Pregnant females were euthanized at E6.5, E7.5/8.5, or E15.5 to count the number of pups per litter and to check for reabsorption sites. Time points were determined based on previous literature indicating Cenpa^−/− embryos are viable until E8.5⁸⁰. Litter sizes and pup genotypes at E15.5 are reported in the table below. Primer sequences for genotyping Cenpa^mScarlet/+ mice are as follows: wildtype allele forward 5’-GGCAGCTTAGGGGAGTCC-3’, mScarlet allele forward 5’-CTGAAAAGAACAGCCAATGTGTAAC-3’, and shared reverse primer 5’-TCCATATGCACTTTAAACCTCATAAATTC-3’.

Microscopy

All fixed mouse samples were imaged on a Nikon A1R HD25 point scanning confocal microscope with a 60X 1.4NA oil objective (Nikon MRD71600). The microscope platform was controlled in Nikon NIS Elements (Version 5.21.03).

Fixed human samples were imaged on a Leica Stellaris 5 point scanning confocal microscope using a 63x objective (HC PL APO 63x/1.40 oil C52).

Immunofluorescence on mouse testis sections

Testes were collected from WT C57BL/6J or Cenpa^mScarlet/+ males between 8–12 weeks of age and snap frozen in liquid nitrogen. The frozen testes were then embedded in OCT medium (Leica 39475237) on dry ice and stored at −80°C. 10μm sections were first dried for 10–20 minutes at RT, fixed in 4% paraformaldehyde (PFA) for 10 minutes, washed with PBS two times for 5 minutes each, then permeabilized in PBS + 0.5% Triton X-100 for 30 minutes at RT. Tissue sections were washed twice in PBS then blocked for 1 hour at RT in filtered PBS + 3% bovine serum albumin (BSA) + 0.1% Triton X-100 before incubation with primary antibodies diluted in filtered PBS + 3% BSA + 0.1% Triton X-100 at 4°C overnight: rabbit α CENP-A at 1:250 dilution (Cell Signaling Technologies 2048), rabbit α CENP-C at 1:2000 dilution (Das et al.³⁰). The next day, tissue sections were washed four times for 15 minutes each at RT in PBS + 0.1% Triton X-100 then labelled with PNA Lectin-FITC (Genetex) diluted 1:1000, DAPI diluted 1:1000, and secondary antibodies diluted 1:1000 in filtered PBS + 3% BSA + 0.1% Triton X-100 for either 1.5–2 hours at RT or overnight at 4°C. Finally, the secondary antibodies were washed out with PBS three times for 10 minutes each and sections were covered in a drop of VectaShield (Vector Laboratories) before mounting.

Immunofluorescence on whole mount testes

Seminiferous tubules from adult Cenpa^mScarlet/+ male mice were prepared as previously published with some modifications⁸⁹. Briefly, the tunica albuginea was removed and seminiferous tubules were teased apart in PBS. Interstitial cells were removed by incubating the tubules in 0.5mg/mL collagenase A1 at RT for 5 minutes. The tubules were then rinsed three times with PBS and fixed in 2% PFA (Electron Microscopy Sciences) for 6 hours at 4°C. Fixed tubules were then rinsed three times with PBS and permeabilized with 0.25% NP-40 in PBST (PBS + 0.05% Tween) for 25 minutes at RT. The tubules were rinsed three times for 5 minutes each with PBST and blocked using 5% normal donkey serum (Jackson Immuno Research) in PBST either 2 hours at RT or overnight at 4°C. The tubules were incubated with one of the following primary antibodies diluted in PBST + 5% normal donkey serum at 4°C overnight: rabbit α CENP-A at 1:250 (Cell Signaling Technologies 2048), goat GFRα at 1:250 (R&D Systems AF429), rabbit Sohlh1 at 1:200 (gift from Aleksandar Rajkovic⁴⁷), rabbit Stra8 at 1:250 (Abcam ab49602), or mouse H2A.X at 1:500 (EMD Millipore 05–636). The next day, the tubules were washed three times for 10 minutes each in PBST, then stained with DAPI at a 1:1000 dilution and secondary antibodies at RT for 1 hour in PBST + 5% normal donkey serum. After secondary antibody incubation, the tubules were rinsed three times for 10 minutes each with PBST, gently spread onto a coverslip with a paintbrush, dried, briefly, and mounted onto slides in Vectashield.

Preparation of spermatocyte spreads

Spermatocytes were mounted and spread onto glass slides as previously described⁹⁰. Briefly, a single testis from a Cenpa^mScarlet/+ male was decapsulated by shaking at 150rpm for 15 minutes at 32°C in 2mL testis isolation medium (TIM) comprised of 104 mM NaCl + 45 mM KCl + 1.2 mM MgSO4 + 0.6 mM KH2PO4 + 0.1% (w/v) glucose + 6 mM sodium lactate + 1 mM sodium pyruvate pH 7.3 with an additional 2mg/mL collagenase A1 added for this step only. The tubules were washed three times in 15mL of TIM at RT before being resuspended in 2 mL of TIM + 0.7 mg/mL Trypsin + 4 ug/mL DNaseI and shaken at 150rpm for 15 minutes at 32°C. The trypsin digest was stopped by adding 1mL of fetal bovine serum (FBS) + DnaseI to 6.67ug/mL in the resulting single cell suspension and cell clumps were separated by gentle pipetting with a plastic transfer pipet and filtering through a 100um nylon mesh filter. After filtering, TIM was added to 15 mL and spun at 200g for 5 min. The cells were then washed twice in TIM + 15uL of 400ug/mL DnaseI and once in 12mL PBS + 15uL of 400ug/mL DnaseI with centrifugation at 200g. The dissociated cells were then separated into twelve equal aliquots and two aliquots were spun at 200g for 5 min spread onto glass slides in turns. Each aliquot was resuspended in pre-warmed to 37°C 80uL 0.1 M sucrose and incubated for 3–5 minutes at RT then 20uL of the aliquot was slowly spread along the length of a positively charged and precleaned glass slide. Each aliquot was spread onto four glass slides, which were each pre-covered with 65uL of prewarmed 1% PFA pH 9.2 + 0.1% Triton X-100 during the aliquot’s 3 to 5-minute incubation in 0.1 M sucrose. After spreading all of the aliquots of the testicular single cell suspension onto glass slides, the slides were allowed to dry slowly at RT for 3.5 hours. Finally, the slides were rinsed one in H₂O and twice for five minutes each in PBS + 0.004% Photo-Flo 200 (Kodak 1464510), air dried, and stored at −80°C.

Immunofluorescence staining of spermatocyte spreads

Spermatocytes spreads were immunostained following a previously published protocol with some modifications⁹⁰. Slides were first dried for 10–20 minutes at RT, then immediately incubated with lambda phosphatase (New England Biolabs, P0753S) for 2 hours at 30°C. Slides were then blocked for 30 minutes at RT in PBS + 1% FBS+ 3% BSA + 0.05% Triton X-100 and incubated with primary antibodies diluted in PBS + 1% FBS + 3% BSA + 0.05% Triton X-100 overnight at RT: mouse α Sycp3 at 1:200 (Abcam ab97672) and rabbit α CENP-A at 1:250 (Cell Signaling Technologies 2048) or rabbit α CENP-C at 1:2000 (Das et al.³⁰). The next day, spreads were washed with PBS + 1% FBS + 3% BSA + 0.05% Triton X-100 four times for 1 minute, 5 minutes, 10 minutes, and 15 minutes, respectively, all at RT. Next, the spreads were stained with DAPI diluted 1:500 and secondary antibodies diluted 1:200 in PBS + 1% FBS + 3% BSA + 0.05% Triton X-100 for 45 minutes at RT. The spreads were then washed again with PBS + 1% FBS + 3% BSA + 0.05% Triton X-100 four times for 1 minute, 5 minutes, 10 minutes, and 15 minutes, respectively, all at RT. All incubations were done in a humified chamber to prevent evaporation. Finally, they were washed twice for five minutes each in PBS + 0.4% Photo-Flo 200, dried briefly, and covered in a drop of VectaShield to mount.

Flow cytometry

Testes from adult Cenpa^mScarlet/+ male mice were dissociated, stained, and flow sorted as previously published with some modifications⁹¹. Briefly, the tunica albuginea was removed and the seminiferous tubules were transferred to 10ml of Advanced DMEM:F12 media + 200ul of 10 mg/ml Collagenase A1 + 100ul of 20mg/mL Dnase I. Tubules were dispersed by shaking at 35C for 5min at 215rpm and allowed to settle for 1 minute at room temperature. Excess media was removed with a sterile pipette, leaving ~2mL with the settled tubules. The tubules were then dissociated at 35°C with horizontal shaking at 215 rpm for 5 minutes in an additional 10mL of Advanced DMEM:F12 media + 200ul of 20mg/ml DnaseI + 5mg of Trypsin The trypsin was quenched with the addition of 3mL of FBS. The resulting single cell suspension was filtered through a 100um strainer washed in PBS pelleted at 300g for 5 minutes, and re-suspended in 6mL MACS buffer containing 0.5% BSA (Miltenyi Biotec). The suspension was stained with Hoechst 33342 and propidium iodide (PI) as previously published with no modifications⁹². Spermatogonia (2n), spermatocytes (4n, mostly pachytene and diplotene), and round spermatids (1n) were isolated from the stained live single cell suspension based on DNA content and DNA compaction using a FACS ARIA II/III flow cytometer (BD Biosciences) and Bigfoot Spectral Cell Sorter available at the University of Michigan’s Flow Cytometry Core. Gates and sorting conditions were adapted from previously published methods and optimized for each biological replicate^91,92. Sorted cell populations were immediately pelleted at 600g and snap frozen in liquid nitrogen before storage at −80°C.

Subcellular and high-salt fractionation of tagged CENP-A testes

Cells were fractionated using a high-salt gradient, as we have published previously⁹³. Briefly, flash frozen testes were homogenized (Dounce homogenizer) and then washed with PBS. Cells was incubated with 0.1% CTAB for 5 min on ice to remove tails from the spermatids and then washed with 50mM Tris-HCl pH 8.0 several times. Cells were lysed for 15 min on ice and then spun down (800 g for 15 min). The supernatant (cytoplasmic fraction) was collected, and the nuclei were digested with 1U of MNase at 37C for 30min. The digestion was quenched with EGTA, then spun down (400g for 10min). The supernatant was collected (MNase fraction) and the remaining sample was fractionated via sequential incubations (30 min each) with varying NaCl concentrations (100 mM, 300 mM, 500 mM, 1 M, and 2 M). Salt fractions were incubated with 20% TCA overnight at −20C to precipitate the proteins and then resuspended in water before immunoblotting. Immunoblotted membranes were incubated with antibodies against CENP-A at 1:250 (Cell Signaling Technologies 2048), V5 at 1:1000 (BioRad MCA1360). Primary antibodies hosted in mouse were detected with HRP-conjugated goat α mouse IgG at 1:10,000 (Abcam ab6721) and primary antibodies hosted in rabbit were detected using HRP conjugated mouse α rabbit IgG light chain at 1:5,000 (Abcam ab99697).

Sperm chromatin immunoprecipitations

Mouse sperm was collected in a swim up from the caudal epididymis and vas deferens as previously described⁹⁴, and mature human sperm was collected and snap frozen as detailed above. Chromatin from both human and mouse sperm was immunoprecipitated using a previously published protocol with some modifications⁹⁵. Briefly, 30 million mature sperm were resuspended in 1mL of PBS and cross-linked by 1% formaldehyde for 10 minutes at RT before quenching the fixation by adding Tris pH 7.5 to a final concentration of 0.2 M and incubating for another 10 minutes at RT. The fixed sperm were then washed twice in sperm decondensation buffer: 5 mM HEPES pH 8.0 + 1 mM PMSF + 0.2% NP-40 + 10 mM EDTA + 5 mM NaCl + 1.2 M urea + 10 mM DTT + 2X complete protease inhibitor cocktail. After washing, the sperm pellet was resuspended in sperm decondensation buffer with 1 mg/mL heparin sodium salt at a concentration of 15 million sperm per 3 mL decondensation buffer. The mouse sperm was then decondensed by incubating for 5 hours at 42°C, while human sperm was decondensed by incubating for 2 hours at 42°C. The decondensed sperm was then washed twice in lysis buffer: 50 mM HEPES pH 7.5 + 140 mM NaCl + 1 mM EDTA + 10% glycerol + 0.5% NP-40 + 0.25% Triton X-100 + and 1X complete protease inhibitor cocktail. After washing, the decondensed mouse sperm was treated with lambda phosphatase by resuspending the sperm in 1X NEBuffer for Protein MetalloPhosphatases supplemented with 1 mM MnCl2 and incubating for 30 minutes at 37°C. The decondensed human sperm was not treated with lambda phosphatase. Both mouse and human sperm were then resuspended in 50 mM Tris-HCl pH 8.0 + 10 mM EDTA + 1% SDS + 1X complete protease inhibitors + 1 mM PMSF and sonicated using four cycles of 30 seconds on/off. The sonicated sperm was diluted 1:10 in 10 mM Tris-HCl pH 8.0 + 100 mM NaCl + 1 mM EDTA + 0.5 mM EGTA + 1% Triton X-100 + 0.1% sodium deoxycholate + 1X complete protease inhibitor cocktail to dilute SDS to a final concentration 0.1%. Next, the sperm chromatin was spun down for 10 minutes at 20,000g and the insoluble fraction was saved for immunoblot analysis. The soluble chromatin was precleared by gently rotating with magnetic Protein A beads (Invitrogen 10002D) for 1 hour at 4°C before pre-clearing beads were removed and 10% of the soluble chromatin was saved as input for immunoblot analysis. Chromatin was immunoprecipitated by incubating the soluble chromatin with the desired antibody pre-bound to magnetic Protein A beads for at least 14 hours at 4°C with gentle rotation. Antibodies used for immunoprecipitation include rabbit α CENP-A at 1:200 (Cell Signaling Technologies 2048) for mouse sperm, rabbit α HJURP at 1 ug per 10 million sperm (Abcam ab100800) for human sperm, and rabbit IgG isotype control at either 1:200 or 1 ug per 10million sperm (Invitrogen 10500C) for mouse or human sperm respectively. The unbound fraction was saved for immunoblot analysis, and the immunoprecipitation was then washed four times in 50 mM HEPES pH 7.0 + 0.5 M LiCl + 1 mM EDTA + 0.7% sodium deoxycholate + 1% NP-40 then twice in 10 mM Tris-HCl pH 8.0 + 1 mM EDTA, and the immunoprecipitated lysates were finally eluted in 10 mM Tris-HCl pH 8.0 + 1 mM EDTA + 1% SDS + 1X complete protease inhibitor cocktail.The input (10% of IP), 10% of the unbound, and 100% of the elution were immunoblotted from each experiment and antibody. Immunoblotted membranes were incubated with either rabbit α HJURP at 1:500 (Abcam ab100800), rabbit α CENP-A at 1:250 (Cell Signaling Technologies 2048), mouse α CENP-A at 1:1000 (GeneTex GTX13939), mouse α H2B at 1:500 (Abcam ab52484), or rabbit α H2B at 1:1000 (Cell Signaling Technologies 12364). Primary antibodies hosted in mouse were detected with HRP-conjugated goat α mouse IgG at 1:10,000 (Abcam ab6721) and primary antibodies hosted in rabbit were detected using HRP conjugated mouse α rabbit IgG light chain at 1:5,000 (Abcam ab99697).

Human sperm immunofluorescence

Ejaculate samples were centrifuged at 2500g for 5 minutes. The supernatant was discarded, and the resulting pellet was resuspended in 500 μL of PBS. Samples were sonicated at 40% power for 20 seconds, rested for 15 seconds, and then sonicated for an additional 5 seconds to remove sperm tails. Following sonication, samples were centrifuged again at 2500g for 5 minutes and the supernatant was discarded. The pellet was resuspended in a Percoll solution composed of 350 μL Percoll and 150 μL DMEM and centrifuged at 2500g for 5 minutes for 5 minutes. The pellet was then resuspended in DMEM, and sperm count was assessed. Samples were centrifuged once more at 2500g for 5 minutes and resuspended in PBS at a concentration of 25,000 sperm/mL. Cells were cytocentrifuged onto microscope slides using a Cytospin device (Hettich Rotofix 32 A). Sperm were permeabilized in 0.5% CTAB in PBS for 30 minutes at RT. Slides were washed in 50 mM Tris buffer (pH 8.0) for 2 minutes. Sperm heads were then decondensed using 25 mM DTT in 500 mM NaCl for 45 minutes. Samples were fixed in 4% PFA for 15 minutes then treated with lambda phosphatase for 2 hours at 30°C. Following treatment, samples were blocked in 3% BSA + 0.1% Triton X-100 in PBS for 1 hour at RT. For immunostaining, samples were incubated overnight at 4°C with rabbit CENP-A antibody (1:250; Cell Signaling, 2186) diluted in blocking solution. Slides were then washed three times in PBS, 10 minutes each wash. Secondary staining was performed using Alexa Fluor 594 goat α rabbit (1:500; Invitrogen A-11012) and DAPI (1:1000) for 2 hours at room temperature. Slides were washed again three times in PBS, 10 minutes each, and mounted with Vectashield.

In vitro fertilization

In vitro fertilization was preformed as described previously⁹⁶ with some modifications. For mouse zygotes and early embryos collected by standard in vitro fertilization (IVF), oocytes collected from either C57BL/6J females (Jax), Cenpa^mScarlet/+, or Cenpa^+/− females. Females were superovulated with an intraperitoneal injection of 100 uL HyperOva (Card KYD-010-EX-X5) and 5–7.5 IU of human chorionic gonadotropin (hCG) (Sigma CG5) at 60 and 14–16 hours respectively before oocyte collection. See “Oocyte collection and maturation” for the isolation of oocytes for in vitro maturation (IVM). Sperm was collected from either B6D2F1/J males (Jax), or Cenpa^mScarlet/+ males. B6D2F1/J and Cenpa^mScarlet/+ males were housed individually for three to seven days prior to IVF.

Media for sperm capacitation and IVF incubation, and if applicable later stage embryo maturation, was allowed to equilibrate at 37°C with 5% CO2 the night before IVF. The day of IVF, males were euthanized 30 minutes to 1.25 hours before oocyte collection. Mature sperm was collected by mincing the epididymis and vas deferens and capacitating in 500 uL Research Vitro Fert Media (Cook Medical K-RVF-50) or Human Tubule Fluid (HTF, Sigma MR-070-D) containing 0.2mg/mL hypotaurine for 10 to 15 minutes at 37°C and 5% CO2. The epididymis and vas deferens tissue were then removed; sperm concentration and percent motility were counted using a Makler sperm counter (Cooper Surgical SEF-MAKL). One million motile sperm were then transferred into a new 500uL well of Research Vitro Fert Media or HTF in preparation for oocytes/cumulus masses.

For standard IVF, females were euthanized 14–16 hours post hCG injection and cumulus-oocyte complexes were extracted from the oviduct in pre-warmed MEM media (ThermoFisher 12360038) and immediately transferred into a well containing 1 million motile sperm. For oocytes generated through IVM, oocytes were removed from MEMa + Glutamax, rinsed at least once through HTF, and transferred to the well of HTF containing sperm. Except for the experiments with drug treatments, zygotes were allowed to incubate with the sperm until the time of collection. For extended embryo culture, fertilization was allowed to continue for 4–6 hours before the zygotes were removed from the well containing sperm and washed 5 times and incubated in pre-equilibrated KSOM AA at 37°C and 5% CO2 and allowed to develop to morula or blastocyst stage embryos.

Inhibiting transcription, translation and DNA replication in zygotes

For inhibitor treatments, fertilization was allowed to continue for 4–6 hours before the zygotes were removed from the well containing sperm and placed in another pre-equilibrated 500 uL of Research Vitro Fert Media or HTF. To inhibit transcription, zygotes were treated with 5,6-dichloro-1-β-D ribofuranosyl-benzimidazole diluted 1:200 to 200 μM and 5-ethynyl uridine (Click Chemistry Tools) diluted 1:500 to 500 nM, both added to the IVF media 1.5 hours after fertilization, embryos were collected 10 hours after fertilization. To block translation, zygotes were treated with cycloheximide diluted 1:500 to 100 μg/mL starting 1 hour after fertilization and O-propargyl-puromycin (Click Chemistry Tools 1493) diluted 1:500 to 20 μM in the IVF media 1 hour before collecting the zygotes 8.5 hours after fertilization. To block DNA replication, zygotes were treated with aphidicolin diluted 1:500 to 10 μg/mL starting 1.5 hours after fertilization and 5-Ethynyl 2'-deoxyuridine diluted 1:500 to 20 μM added 1 hour before collection at 8.5 hours after fertilization. For all drug-treatment experiments, control zygotes were fertilized and cultured in parallel and subjected to identical labeling (EU, OPP, or EdU), except that equivalent volumes of dimethyl sulfoxide were added to the IVF medium in place of the small-molecule inhibitors.

Modulating cell cycle in zygotes

For CDK1/2 inhibition, zygotes were treated with 5 μM flavopiridol (Selleckchem); for PLK1 inhibition, 10 nM BI2536 (Selleckchem) was used. Inhibitors or vehicle (0.1% DMSO) were added directly to the IVF medium 1.5 h after fertilization, and embryos were cultured under these conditions until collection at 8.5 h post-fertilization. Control embryos were handled in parallel with vehicle only. The timing and method of compound addition followed the same protocol described above for DRB, cycloheximide, and aphidicolin treatments.

Immunofluorescence on oocytes and early embryos

Embryos or oocytes were collected at the indicated timepoints after washing in PBS, treated briefly (30 seconds-1 minute) with Acidic Tyrodes solution to remove the zona pellucida, and fixed in 4% PFA + PBS + 0.04% Triton X-100 + 0.3% Tween-20 + 0.2% sucrose for 10–15 minutes at 37°C. Fixed embryos or oocytes were stored under mineral oil in PBS at 4°C for no longer than two weeks. All staining took place in nine-well depression Pyrex plates (Millipore Sigma CLS722085) and utilized buffers which were made fresh and filtered through 0.45 um PES filters (Whatman 6780–2504). On the first day, embryos or oocytes were permeabilized in PBS + 0.5% Triton X-100 + 3% BSA at RT for 1 hour. If the embryos were incubated with EdU, 5-EU, or OPP (see above), these reagents, which each contain a terminal alkyne group, were next conjugated to an AZDye 488 picolyl azide following the Click Chemistry Tools kit protocol (CCT #1493). If no additional immunostaining was necessary the treated embryos were stained with DAPI diluted at 1:200 in PBS + 0.1% Triton + 3% BSA for 15–30 minutes at room temperature, washed twice for 10 minutes each with PBS + 0.1% Triton + 3% BSA, briefly dried onto a glass slide, then covered in Vectashield. After permeabilization and optional click-chemistry staining and/or phosphatase treatment (see above), embryos or oocytes were blocked in PBS + 0.1% Triton + 3% BSA + 10% FBS (blocking buffer) at RT for 1–2 hours and stained with primary antibodies diluted in blocking buffer at 4°C overnight: rabbit α CENP-A at 1:250 dilution (Cell Signaling Technologies 2048), rabbit α Mis18BP1 at 1:100 (Sigma HPA006504), rabbit α CENP-C at 1:1000 dilution (Das et al.³⁰), or rabbit α CENP-C at 1:2000 dilution (gift from Iain Cheeseman⁹⁷). Embryos or oocytes were washed the following day in PBS with 0.1% Triton and 3% BSA five times for at least 15 minutes each, followed by incubation with DAPI at 1:200 and secondary antibodies diluted 1:500 in PBS + 0.1% Triton + 3% BSA + 10% FBS for 2 hours at RT or overnight at 4°C. The stained embryos or oocytes were washed thrice for 10 minutes each with PBS + 0.1% Triton + 3% BSA, briefly dried onto a glass slide, then covered in Vectashield. Staining strategy was also used for human oocytes.

Oocyte collection and maturation

Oocytes were collected and matured in vitro as described previously⁹⁸. Germinal vesicle (GV) stage mouse oocytes were collected from either CF-1 females (Charles River Laboratories), C57BL/6J females (Jax), or Cenpa^mScarlet/+ at 3–4 or 6–16 weeks of age. The females were each injected with 5–7.5 IU pregnant mare’s serum gonadotropin or 100 uL HyperOva intraperitoneally 44–48 hours prior to sacrifice. MEMa with GlutaMAX (ThermoFisher 32561037) medium for oocyte collection and maturation were placed in an incubator to equilibrate to 37°C and 5% CO2 the day before oocyte collection. Milrinone (Sigma, M4659–10mg) and FBS were diluted to 2.5 μM and 5% respectively in both oocyte and collection media about an hour before oocyte collection. Females were euthanized and ovaries were dissected out and placed in pre-warmed MEM (ThermoFisher 12360038) + 2.5 μM milrinone + 5% FBS + 3mg/mL Polyvinylpyrrolidone (Sigma, PVP10–100G). Oocyte-cumulus cell complexes were isolated by repeatedly poking the ovaries’ antral follicles with an insulin syringe needle. Oocytes were washed through several drops of MEM + 2.5 μM milrinone + 5% FBS and cumulus cells were mechanically detached by repeated mouth pipetting the complexes up and down. Cleaned GV oocytes were then washed and allowed to recover in MEMa with GlutaMAX (ThermoFisher 32561037) + 2.5 μM milrinone + 5% FBS for several hours at 37°C and 5% CO2. For maturation, the GV oocytes were washed away from milrinone by transferring them through at least 5 drops of MEMa with GlutaMAX (ThermoFisher 32561037) + 5% FBS, at which point they were then cultured at 37°C and 5% CO2. Metaphase I (MI) oocytes were collected 8 hours after release from milrinone and metaphase II (MII) oocytes were collected 16 hours after release from milrinone.

Oocyte microinjection

GV oocytes were collected as described above and microinjected following a previously published protocol⁹⁸. Briefly, after letting GV oocytes recover from collection, the oocytes were transferred into MEM (ThermoFisher 12360038) + 2.5 μM milrinone (Sigma) + 5% FBS and injected in groups of 20–30 with a FemtoJet 4i microinjector (Eppendorf). Holding micropipettes (Cooper Surgical MPH-MED-35) were cleaned with 70% ethanol and ddH₂O, and injecting needles (Sutter B100–75-10) were pulled with a Sutter P97 or Sutter P2000 micropipette puller fresh before every injection. Injection pressure was adjusted for each needle, and injection volumes were approximately the same size as the GV (typically around 500 fPa).

Pools of siRNAs were purchased from Horizon Discovery and reconstituted in RNase-free dH₂O according to the product manual. siRNA pools were stored in aliquots at −80°C. On the day of injection, individual aliquots were thawed on ice, diluted to the appropriate concentration with RNase-free dH₂O, and spun down at 20,000g for 30min at 4°C. Pools used for this study are: CENP-A (J-044345–05-0002 through J-044345–08-0002, combined or L-044345–00-0005 pre-pooled), CENP-C (L-044347–00-0005), and negative control (D-001810-10-05).

In vitro transcribed RNA was prepared for injection by either linearizing the pIVT:: CENP-B-EGFP plasmid with PvuI (NEB) or double digesting pGEMHE::CENP-C plasmids with SbfI and PvuI (both NEB). 100–1000ng of linearized plasmid was then utilized as the template for a T7 in vitro transcription reaction with the mMessage mMachine T7 kit (ThermoFisher AM1344), depending on the construct. RNA was purified with Qiagen’s RNeasy MinElute Cleanup Kit (74204) or with phenol chloroform per the instructions in the mMESSAGE mMachine T7 kit, aliquoted, and stored at −80°C. While the collected GV oocytes were recovering, RNA integrity was checked by denaturing gel electrophoresis and good quality IVT RNA was combined, if necessary, with siRNA pools. RNA was diluted to reported concentrations in RNase-free dH₂O, and spun down at 20,000g for 30min at 4C. Injected GV oocytes were allowed to recover in MEMa with GlutaMAX + 2.5 μM milrinone + 5% FBS for 1–6 hours at 37C and 5% CO₂. The injected oocytes were matured and fertilized as described above.

siRNA-resistance was conferred in Cenpc^WT IVT plasmid by introducing synonymous mutations within the seed regions for each siRNA in the targeting pool. The Cenpc^Dimer IVT was generated from the Cenpc^WT IVT plasmid, with point mutations derived from previous literature introduced into the corresponding mouse sequence residues (F829A and F901)^37,70.

Oocyte photobleaching

GV oocytes were collected as described above and imaged and photobleached according to previously published protocols^99,100. For photobleaching, after recovering from collection, the GV oocytes were stained with 10ug/mL Hoechst 33324 (Sigma) for 30 minutes. Stained oocytes were then incubated in droplets of MEMa with GlutaMAX (ThermoFisher 32561037) + 2.5 μM milrinone + 5% FBS pre-equilibrated to 37°C at 5% CO2 under oil (NidOil, NO-400K). Droplets of oocytes were imaged and bleached in a MatTek 35mm, 1.5 coverslip, 10mm glass bottom dish (P35GC-1.5–10-C). The GV oocytes were incubated in an OkoLab live-cell imaging chamber (H301) on a Nikon Nikon A1R with a 60X 1.2NA long range water-immersion objective available at the University of Michigan’s Center for Live Cell Imaging through the Microscopy and Image Analysis Laboratory. Nikon NIS Elements was used to photobleach the mScarlet fluorescence in the nucleus of each GV oocyte with a circular ROI, the area adjusted to the size of each nucleus, with a 561nm laser line at 90μW of power and a 19.9μSec pixel dwell time. After photobleaching, the oocytes were allowed to recover for at least 30 minutes before being in vitro matured and subsequently in vitro fertilized as described above. All zygotes from photobleached, in vitro matured, and in vitro fertilized oocytes were collected after 8.5 hours of fertilization.

Generation of androgenetic zygotes by intracytoplasmic sperm injection (ICSI)

Ovulation was induced by intraperitoneal injection (i.p.) of equine chorionic gonadotropin (Pregmagon, IDT) and human chorionic gonadotropin (Ovogest, Intergonan). Lean B6C3F1 females aged 8 weeks and weighing approx. 25 g were injected i.p. using a 27G needle, at 5pm, with 10 I.U. eCG and 10 I.U. hCG 48 h apart.

Cumulus-oocyte complexes were collected from oviducts 14 hours after hCG. The cumulus cells were dispersed in 50 U/mL hyaluronidase dissolved in Hepes-buffered Chatot, Ziomek and Bavister medium (HCZB)¹⁰¹ with BSA replaced by polyvinylpyrrolidone (PVP, 40 kDa) 0.1% w/v, at room temperature (28 °C). The cumulus-free MII oocytes were cultured in 500 μL of α-MEM medium (Sigma, M4526) supplemented with 0.2% (w/v) BSA and 50 mg/mL gentamicin, in 4-well plates (Nunc), in a 6% CO 2 incubator at 37 °C until further processing.

Diploid androgenetic embryos were generated by stripping the oocytes of their meiotic spindle followed by injection of two sperm heads, on the stage of a NikonTE2000 inverted microscope fitted with Nomarski optics, at room temperature (28 °C). To remove the meiotic spindle, oocytes were placed in a 300 μL elongated drop of HCZB medium supplemented with 0.1% PVP and 5 μM Latrunculin B (428020, Merck Millipore, Darmstadt, Germany), on a glass-bottomed dish (Figure 2). A hole was drilled in the zona pellucida using a mercury-filled, piezo-driven needle (inner diameter 10–12 μm), and the spindle was gently aspirated. Resultant ooplasts were allowed to recover in α-MEM medium for at least 2 hours. For ICSI, ooplasts were placed in a 300 μL elongated drop of HCZB medium supplemented with 0.5 % (w/v) PVP on a glass-bottomed dish as described. Each ooplast was injected with two sperm heads simultaneously using a mercury-filled, piezo-driven needle (inner diameter 7–8 μm). Ten minutes after the ICSI the oocytes were transferred to 500 mL of Potassium (K) simplex optimization medium (KSOM) in 4-well plates, in a 6% CO2 incubator at 37 °C. KSOM was prepared in house as per the original recipe¹⁰² and contained free amino acids both essential and non-essential, 0.2% (w/v) BSA and gentamicin (50 mg/mL). Controls for the diploid androgenetic embryos were generated by ICSI of single sperm heads in intact oocytes. Both groups were cultured in KSOM medium to allow for oocyte activation and pronuclear formation, until further processing.

After 16 hours of cultures, the embryos were fixed in 3.7% PFA for 1 hour at room temperature, permeabilized in 0.5% Triton X-100 for 20 minutes, and then blocked/stored for transport 2% BSA + 2% glycine + 0.1% Tween-20 + 5% Donkey Serum. Subsequent immunofluorescence staining was performed as described above, beginning at the phosphatase treatment.

Representative Images

Representative images were acquired primary using Fiji/ImageJ software to create maximum intensity projections using selective z-slices from zygotes to collapse both pronuclei into one image. Additional images were acquired using ImarisViewer 10.1.0. Images were imported and opened in the “Section” view with extended settings to create maximum intensity projections.

Representative Video

A video demonstrating the identification and quantification of puncta was recorded using Apple's built-in screen recording feature on a device running iOS 18.4. The recorded footage was subsequently edited using iMovie version 10.4.3

QUANTIFICATION AND STATISTICAL ANALYSIS

Quantification of centromeric CENP-A and CENP-C signals

Quantification of centromere fluorescence intensity was made in Fiji/ImageJ software by drawing circles of constant diameter around individual fluorescent puncta associated with DNA stains across all Z-planes of the cell. The total intensity was calculated for each punctum after subtracting the background of the image, which was obtained from un-stained regions of the cell nucleus for each cell and then summing the total puncta fluorescence intensities for each pronuclei or cell being studied (see Video S1). For each statistical test, multiple independent technical and biological replicates were quantified to obtain a minimum of about ten embryos per experiment.

Quantification of the total CID signal per gamete was performed following a previously established method^88,103. Briefly, un-deconvolved raw images were collected as 2D Z-stacks and saved as un-scaled 16-bit TIFF files. Fluorescence signals were manually quantified using Fiji (ImageJ). A circular region of interest (ROI) was drawn around the entire fluorescence signal for each gamete (either CID-Dendra2 or Hoechst signal). An identical-sized circle was drawn in an adjacent background area. The raw integrated density (RawIntDen), representing the sum of pixel gray values within each ROI, was determined for both signal and background regions. The net fluorescence signal (Foreground signal) was calculated by subtracting the background signal from the raw signal. The total fluorescence per gamete was then determined by summing the background-corrected fluorescence from all Z-stacks. All CID quantifications were conducted using this approach.

Quantification of MIS18BP1 staining in zygotes

Quantification of centromere fluorescence intensity was made in Fiji/ImageJ software by drawing circles around the maternal and paternal pronuclei at the z-slice in which each pronucleus was largest. Raw integrated intensity measurements are normalized by pronuclear area and plotted as the paternal/maternal ratio where reported.

Quantification of mis-segregated chromosomes

For 2-cell embryos that were scored for mis-segregated chromosomes, embryos were collected 24hpf to allow sufficient time for nuclear envelope formation to occur in both blastomeres. Mis-segregated chromosome counts were determined by counting the number of DAPI-stained structures outside of the nuclear envelopes in both blastomeres. Individual chromosome counts were not determined.

Statistical testing

All statistical tests for significance were performed in either R or GraphPad Prism 9 or 10 software. P-values were calculated at a significance level of 0.05 or 95% confidence interval for all two tailed statistical tests. Figures were made in R and GraphPad Prism 9 software. The number of replicates for each experiment is supplied in the figure legends. Tests for normal distribution of the data were performed in R using both the shapiro.test() and leveneTest() functions. Normally distributed paired data was tested using the t.test() and non-parametric data was tested using the wilcox.test(). In GraphPad, the Shapiro-Wilk normality test was applied to determine if the data was normally distributed, and the appropriate normal or non-parametric t-test was applied to the data accordingly.

Statistical testing for 2-cell developmental progression and 2-cell mis-segregation rates were performed in R. P-values were calculated at a significance level of 0.05 using the prop.test() function.

Supplementary Material

2

Video S1: Demonstration of centromeric signal quantification, related to Methods.

3

Table S1: Cenpa^mScarlet/+ intercrosses produce reduced litter sizes by E15.5, related to Figure 1.

4

Table S2: Cenpa^{mScarlet/mScarlet} pups are not recovered from Cenpa^mScarlet/+ intercrosses, related to Figure 1.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-CENP-A (mouse)	Cell Signaling Technologies	Cat#2048; RRID:AB_1147629
Rabbit anti-CENP-C (mouse)	Das et al.30	N/A
Goat anti-GFRα	R&D Systems	Cat#AF429; RRID:AB_2294621
Rabbit anti-Sohlh1	Gift from Aleksandar Rajkovic47	N/A
Rabbit anti-Stra8	Abcam	Cat#ab49602; RRID:AB_945678
Mouse anti-γH2A.x	Sigma	Cat#05-636; RRID:AB_309864
Mouse anti-Sycp3	Abcam	Cat# ab97672; RRID:AB_10678841
Mouse anti-V5	BioRad	Cat# MCA1360; RRID:AB_322378
Rabbit anti-HJURP	Abcam	Cat# ab100800; RRID:AB_10672568
Rabbit IgG isotype control	ThermoFisher	Cat#10500C; RRID:AB_2532981
Mouse anti-CENP-A (human)	GeneTex	Cat# GTX13939; RRID:AB_369391
Mouse anti-H2B	Abcam	Cat# ab52484; RRID:AB_1139809
Rabbit anti-H2B	Cell Signaling Technologies	Cat#12364; RRID:AB_2714167
Rabbit anti-CENP-A (human)	Cell Signaling Technologies	Cat#2186; RRID:AB_10828491
Rabbit anti-Mis18BP1	Sigma	Cat# HPA006504
Rabbit anti-CENP-C (mouse)	Gift from Iain Cheeseman97	N/A
PNA Lectin-FITC	GeneTex	Cat# GTX01508
DAPI	Sigma	Cat# D9542-1MG
Alexa Fluor 647 donkey anti-rabbit	JacksonImmuno	Cat# 711-605-152; RRID: AB_2492288
Alexa Fluor 488 donkey anti-goat	ThermoFisher	Cat#A-11055; RRID:AB_2534102
Alexa Fluor 488 goat anti-mouse	ThermoFisher	Cat#A28175; RRID:AB_2536161
Alexa Fluor 568 goat anti-mouse	ThermoFisher	Cat#A-11004; RRID:AB_2534072
Alexa Fluor 647 goat anti-mouse	ThermoFisher	Cat#A-21235; RRID:AB_2535804
Alexa Fluor 488 donkey anti-rabbit	ThermoFisher	Cat# A-21206; RRID:AB_2535792
HRP-conjugated mouse anti-rabbit IgG	Abcam	Cat# ab99697; RRID:AB_10673897
HRP-conjugated goat anti-mouse IgG	Abcam	Cat# ab6721; RRID:AB_955447
Alexa Fluor 594 goat anti-rabbit	ThermoFisher	Cat# A-11012; RRID:AB_2534079
Alexa Fluor 568 goat anti-rabbit	ThermoFisher	Cat# A-11011; RRID:AB_143157
Bacterial and virus strains
Biological samples
Human semen samples	University of Michigan Center for Reproductive Medicine	N/A
Human oocytes	University of Michigan Center for Reproductive Medicine	N/A
Chemicals, peptides, and recombinant proteins
NP-40	US Biologicals	Cat# 9036-19-5
DNaseI	Sigma	D4527-200KU
Collagenase	Sigma	Cat# C9891-100MG
Lambda Phosphatase	New England Biolabs	Cat# P0753S
Propidium Iodide	Alfa Aesar	Cat#J66584
CTAB	EMD Millipore	Cat#219374
TCA	Sigma	Cat# T0699-100ML
PMSF	Sigma	Cat# 78830-1G
DTT	Sigma	Cat# 10197777001
Complete protease inhibitor cocktail	Sigma	Cat# 11836170001
Percoll	Sigma	Cat# P1644
5,6-dichloro-1-β-D ribofuranosylbenzimidazole	Sigma	Cat#287891
5-ethynyl uridine	Click Chemistry Tools	Cat#1261-10
Cycloheximide	Sigma	Cat# C7698
O-propargyl-puromycin	Click Chemistry Tools	Cat#1493
aphidicolin	Sigma	Cat# A0781
5-Ethynyl 2'-deoxyuridine	Click Chemistry Tools	Cat#1149-100
Acid Tyrodes	Sigma	Cat#MR-004-D
Milrinone	Sigma	Cat# M4659-10mg
Polyvinylpyrrolidone	Sigma	Cat# PVP10-100G
Hyaluronidase	ICN Biomedicals	Cat#151271
Latrunculin B	Merck Millipore	Cat#428020
Flavopiridol	Selleckchem	Cat#S2679
BI2536	Selleckchem	Cat#S1109
Hypotaurine	Sigma	Cat#H1384-250MG
Hoechst 33342	Thermo Fisher	Cat#H3570
Critical commercial assays
Click & Go Plus 488 Imaging Kit	Click Chemistry Tools	Cat# CCT-1314
mMESSAGE mMachine T7	Thermo Fisher	Cat# AM1344
RNeasy MinElute Cleanup Kit	Qiagen	Cat#74204
Deposited data
Raw western blots	This paper; Mendeley Data	doi: 10.17632/yr4vh7pds7.1
Experimental models: Cell lines
Experimental models: Organisms/strains
CenpamScarlet/+ mouse line	This paper	N/A
Cenpa+/− mouse line	Smoak et al.29	N/A
C57Bl/6J mouse strain	Jackson Laboratories	Cat#000664; RRID:IMSR_JAX:000664
CF-1 mouse strain	Charles River Breeding Labs	Cat#023; RRID:IMSR_CRL:023
B6D2F1/J mouse strain	Jackson Laboratories	Cat# 100006; RRID:IMSR_JAX:100006
B6C3F1 mouse stain	In-house	N/A
CID-Dendra2 Drosophila strain	Ranjan et al.88	N/A
Oligonucleotides
CENP-A siRNA pool	Horizon Discovery	Cat# L-044345-00-0005
CENP-C siRNA pool	Horizon Discovery	Cat# L-044347-00-0005
CENP-A siRNA	Horizon Discovery	Cat# J-044345-05-0002
CENP-A siRNA	Horizon Discovery	Cat# J-044345-06-0002
CENP-A siRNA	Horizon Discovery	Cat# J-044345-07-0002
CENP-A siRNA	Horizon Discovery	Cat# J-044345-08-0002
Non-targeting siRNA pool	Horizon Discovery	Cat# D-001810-10-05
WT Cenpa allele forward primer 5’-GGCAGCTTAGGGGAGTCC-3’	This paper	N/A
mScarlet Cenpa allele forward primer 5’-CTGAAAAGAACAGCCAATGTGTAAC-3’	This paper	N/A
Shared Cenpa allele reverse primer 5’-TCCATATGCACTTTAAACCTCATAAATTC-3’.	This paper	N/A
Recombinant DNA
pIVT:: CENP-B-EGFP	This paper	N/A
pGEMHE:: CENP-CWT-EGFP	This paper	N/A
pGEMHE:: CENP-CDimer-EGFP	This paper	N/A
Software and algorithms
Fiji/ImageJ	https://imagei.net/software/fiii/downloads	RRID:SCR_003070
ImarisViewer	https://imaris.oxinst.com/imaris-viewer	N/A
R Studio	https://posit.co/download/rstudio-desktop/	N/A
Other
VectaShield	Vector Laboratories	Cat#H-1000-10
Photo-Flo 200	Kodak	Cat#1464510
Normal Donkey Serum	Jackson Immuno Research	Cat# 017-000-121
MACS buffer	Miltenyi Biotec	Cat#130-091-222
Protein A Beads	Invitrogen	Cat# 10002D
HyperOva	Cosmo Bio USA	Cat# KYD-010-EX-X25
Human chorionic gonadotropin	Sigma	Cat# CG5-1VL
Research Vitro Fert Media	Cook Medical	Cat# K-RVF-50
Human Tubule Fluid	Sigma	Cat# MR-070-D
KSOM AA	Sigma	Cat# MR-101-D
NidOil	IVF store	Cat# NO-400K
MatTek 35mm, 1.5 coverslip, 10mm glass bottom dish	MatTek	Cat# P35GC-1.5-10-C
Equine chorionic gonadotropin	Ceva Tiergesundheit	N/A

Highlights:

• CENP-A asymmetry between mature gametes is conserved across flies, mice, and humans

• Parental asymmetry resolves before zygotic mitosis via maternal, cytoplasmic CENP-A

• Zygotic CENP-A levels are regulated in a pronucleus-autonomous manner

• CENP-A equalization relies on asymmetric CENP-C recruitment to the paternal pronucleus